Note FYI: I also added a modifier +rrhor to -D. The reason is as follows:

I have a variable density grid used with topography to drive flexure. This is done via -Hrho.grd. However, the density structure rho(x,y) is not perfectly known and I will be running numerical experiments to modify the rho.grd grid by a small scalar 1+f (with f in the ±0.2 range and determine the f that minimizes some rms). So, multiplying rho(x,y) by f changes the mean load density. When the infill density and the load density are set separately then we do the Wessel [2001;2016] approximation which requires these two densities, and now the rho load density will change. However, I wish to keep the same root density (and infill density) in all these experiments, hence I wish to pass that constant root density and the scaled rho.grd so that the approximation stays the same by just adjusting the topography by 1+f. That way the flexural response to me changing 1+f is given by (1+f) times w(x,y).

Anyway, I suspect I am the only one doing this, and it does not affect any tests.